For me, the largest concern with a "personal context" is privacy. Siri is prone to accidental waking, so the addition of your current screen being sent off to the cloud on a false activation, with all the associated data hoovered in training risks, is a place I don't want to be. And yes, i know Siri is processed in the cloud, but it's not currently getting all that context bundled in/
Siri is dumb as a bucket of hair, but for simple tasks like "set a timer" and "what's the news" it is fine 80% of the time (about the same reliability of the top performing LLMs except Siri only 'finds this for you on the web' instead of the LLM making the made up facts sound plausible).
I imagine it will go behind a switch, but if we look how much of "Apple Intelligence" still remains while it's supposed to be turned off, so just on the current state, I do fear that even with Siri turned off, the screen context will still eb sent to some AI company and leaked.
I really hope Apple pull back from this, and differentiate themselves from Google and Microsoft. People who wish to interact with these markov chains, can already install one of their existing apps.
Siri is dumb as a bucket of hair, but for simple tasks like "set a timer" and "what's the news" it is fine 80% of the time (about the same reliability of the top performing LLMs except Siri only 'finds this for you on the web' instead of the LLM making the made up facts sound plausible).
I imagine it will go behind a switch, but if we look how much of "Apple Intelligence" still remains while it's supposed to be turned off, so just on the current state, I do fear that even with Siri turned off, the screen context will still eb sent to some AI company and leaked.
I really hope Apple pull back from this, and differentiate themselves from Google and Microsoft. People who wish to interact with these markov chains, can already install one of their existing apps.